gms | German Medical Science

Objective: Prognosis of adult patients with WHO grade II-III gliomas is not accurately assessed by conventional histological classification. Recently published studies suggest improved classification on the basis of three molecular markers, namely IDH- and TERT-mutational status, and 1p/19q co-deletion status. TERT-mutational status has been shown to gain unfavourable prognostic relevance particularly in IDH wild-type tumors. We here prospectively analysed whether dynamic ¹⁸F-FET data could add to the prognostic profile of suspected low-grade gliomas independent from molecular biomarker.

Method: Time activity curves (TAC) of dynamic PET were classified as homogeneous increasing, focal decreasing or homogeneous decreasing throughout the entire tumor volume. The status of molecular biomarkers was determined using established molecular techniques. TERT-mutational status was analyzed in IDH wild-type tumors. Progression free survival (PFS) and overall survival (OS) were analyzed with the Kaplan-Meier method and prognostic factors were obtained from multivariate regression models. Informed consent was obtained from all patients.

Results: 98 patients with non-enhancing suspected low-grade tumors were analyzed with a median follow-up of 72 months [39-107]. Conventional histological classification revealed 54 low- and 44 high-grade gliomas. Wild-type IDH with/-out TERT-mutation was seen in 22/8 tumors. IDH mutated tumors with/-out 1p/19q co-deletion occurred in 28/40 tumors. IDH wild-type tumors did worst particularly in case of an additional TERT mutation (2-/5-years OS 80% / 40% vs 38% / 6%, p<0.01), whereas 1p/19q co-deleted tumors did best (92% / 85%; p<0.001). Outcome of IDH mutated tumors without 1p/19q co-deletion was in-between (95% / 76%). Increasing or focal decreasing TAC patterns were predominantly seen in IDH mutated tumors lacking 1p/19q co-deletion (80%), while those with wild-type IDH mostly experienced a homogeneous decreasing TAC (66%). This correlation was highly significant (p<0.0001). TAC pattern in 1p/19q co-deleted tumors was highly variable. COX regression analyses indicated a powerful influence of both, the molecular profiling (p<0.001) and the TAC patterns (p<0.001) on OS and PFS.

Conclusions: In suspected low-grade gliomas, the analyses of TAC patterns provide a powerful imaging derived biomarker even after the adjustment for the effects of so far known molecular biomarkers. Evaluation of TAC patterns improves non-invasive prognostic evaluation of these tumors.